In this study, we present a novel MAGL PET radioligand, [¹¹C]PF-06809247 which showed high uptake in the NHP brain and clear blocking effect by a selective MAGL inhibitor. The uptake of [¹¹C]PF-06809247 was relativley uniform thoughout the brain with slightly higher uptake in the celleberum and putamen and lower uptake in the cerebral cortical regions. The distribution was relatively similar as MAGL mRNA expression in the rat brain previously reported [16]. Brain uptake was clearly decreased in all brain regions following pretreatment with a high dose of a previously described selective MAGL inhibitor. The relationship between plasma concentration and occupancy was observed in a dose-dependent manner. This suggests that [¹¹C]PF-06809247 is sensitive enough to detect a range of MAGL inhibition and demonstrates the ammenability of [¹¹C]PF-06809247 for assessing MAGL occupancy in vivo.

The peak brain uptake was higher at a pretreatment condition than baseline although the decreasing was rapid. Compared to baseline condition, the peak radioactivity of the plasma input function after pretreatment was around 1.8 times higher on average. (Additional file 1: Figure S3) That was one possible explanation for increasing brain uptake. It might be due to peripheral blocking by pretreatment drug although there was no data about whole body imaging of pretreatment condition.

The radio-HPLC data showed that [¹¹C]PF-06809247 was rapidly metabolized to less than 10% of total radioactivity in plasma by 30 min. The major radiometabolite could be separated by radio-HPLC and had a retention time that was short (1.9 min) compared to parent fraction (3.1 min), indicating a low probability of brain penetration by the radiometabolite. The metabolisms of PET radioligands among the species such as rodent, NHP and human often show large variabilities. This factor affects the kinetics in the brain according to the change of plasma input function [17]. Despite relatively similar metabolism between human and NHP, actual human studies are needed for accurate evaluations. Additionally, the metabolism of PET radioligand in the brain should be considered. Because of the lack of available data for this issue, further evaluetion also will be needed.

The 2TC model with 3-parameters using arterial input function could describe the TACs of [¹¹C]PF-06809247. A 2TC model with 4-parameters did not descirbe the data well (data not shown). Also, the ratio of brain uptake to metabolite-corrected plasma showed continuously increasing at the duration of quantification (Additional file 1: Figure S4). These imply that [¹¹C]PF-06809247 has irreversible binding properties and is indeed consistent with prior pharmacological study [8]. The Patlak method was also well fitted (Additional file 1: Figure S5), and the Patlak slope was well correlated with K_i by 2TC model estimation although a slight underestimation (around 7%) was observed. This suggests that graphical analysis would be useful for further evaluations such as parametric image analysis.

The whole-body measurement showed highest accumulation in the liver and small intestine, and relatively low accumulation in the urinary bladder. Notably, the uptake in the brain was also high (7% ID at peak). The effective dose of [¹¹C]PF-06809247 was 4.3 μSv/MBq, which was similar to the median value (4.7 μSv/MBq) of other carbon-11 PET radioligands [18]. This radiation exposure would allow multiple administrations in a single subject enabling longitudinal and/or baseline blocking studies in human.

Recently, several studies for NHP brain imaging using PET radioligands for MAGL were reported using [¹¹C]MAGL-0519 [19], [¹¹C]SAR127303 [20], [¹¹C]MA-PB-1 [21], [¹⁸F]T-401 [22], and [¹⁸F]PF-06795071 [1]. In addition, a report highlighting the opportunities and challenges of PET imaging of the endocannabinoid system was recently published [23]. Compared to these PET radioligands, [¹¹C]PF-06809247 showed similar or higher brain uptake (4 SUV at peak) with clear blocking effect by a defined MAGL inhibitor. This data suggests that [¹¹C]PF-06809247 would be a useful PET radioligand for human studies.

One potential limitation of this study is in the use of K_i as the primary outcome measure for determining target occupancy. Theoretically, the rate constant k₃ would be more appropriate, under the conventional assumption that its value is proportional to the available enzyme concentration. However, as shown in Table 2, the estimated k₃ values were highly variable, so this outcome measure was not appropriate for occupancy estimates. This behavior is quite common for irreversibly bound PET radioligands, especially under baseline conditions where enzyme availability is high. Specifically, if baseline k₃ is much higher than k₂ (the clearance rate from tissue), radioligand uptake is flow limited [24]. Thus, using K_i as the outcome measure will likely lead to underestimation of baseline enzyme availability. Under blocking conditions, k₃ is reduced and enzyme availability becomes the rate limiting step, in which case K_i will be more reflective of remaining enzyme availability, i.e., no longer flow limited. It is thus possible that occupancy was underestimated (since baseline K_i is underestimated), in which case the EC₅₀ may be overestimated.